The intention to adopt teleworking: A study in the
healthcare industry at Lentis
Robert F de Wit Student number: S1643703
W A Scholtenstraat 12-53 9712 KW Groningen, The Netherlands
Phone: +31 6 20711766 E-mail: robertdewit.1@gmail.com
The intention to adopt teleworking: A study in the
healthcare industry at Lentis
ABSTRACT
Healthcare organizations have an emerging need for efficiency. The trend of shortages of personnel in the healthcare sector is continuing, because of a lower supply of employees and an increasing demand on healthcare, which can cause future problems. Therefore, organizations in the healthcare industry are trying to find ways to avoid or limit these problems. In mostly commercial organizations, teleworking is already used as a solution to boost flexibility and ultimately efficiency.
In this research, factors influencing the intention to adopt teleworking at the individual level are examined. Contributions to both adoption and teleworking literature are made. This research was performed at the mental care organization “Lentis”. Based on a detailed literature review of adoption models, teleworking factors, and interviews, a research model on the basis of an extended Technology Acceptance Model (TAM) with original variables perceived ease of use, perceived usefulness, attitude, intention to use, and external variables subjective norm, perceived flexibility benefits, personal innovativeness, empowerment and trust was used to perform this research. Both qualitative as quantitative research was conducted to get a better understanding of the proposed relationships. A quantitative questionnaire was used, which was completed by 93 interviewees. The hypotheses were tested using linear regression analysis. Results showed that perceived ease of use was the most important determinant for the attitude toward using teleworking. Perceived usefulness, the subjective norm, personal innovativeness and the perceived flexibility benefits were also proven to influence the attitude toward using teleworking, while the findings for both trust and empowerment were not significant. In this research, the implications of these findings for practice and theory are discussed further. Finally, the research model has an explanatory power of 70%, which was found to be higher than the explanatory power of the TAM constructs. This finding supports that telework adoption cannot only be explained by technological factors.
Contents
1. Introduction ... 4 2. Literature review ... 7 2.1 Teleworking ... 7 2.1.1 Technology ... 7 2.1.2 Work organization ... 7 2.1.3 Social behaviours ... 8 2.1.4 Culture ... 92.2 Behavioral adoption models ... 9
2.2.1 Technology acceptance model (TAM) ... 9
2.2.2 Theory of planned behavior (TPB) ... 10
2.2.3 Unified Theory of Acceptance and Use of Technology (UTAUT) ... 12
2.3 Decision for an adoption model ... 14
2.4 Initial interviews ... 16
3. Research model ... 19
3.1 TAM components ... 19
3.1.1 Technology ... 19
3.2 Extended model based on teleworking features and initial interviews ... 20
3.2.1 Work organization ... 20 3.2.2 Social behaviours ... 21 3.2.3 Culture ... 21 3.2.4 Personal characteristics ... 21 3.3 Control variables ... 22 4. Research methodology ... 23 4.1 Measures ... 23 4.2 Sample... 24 4.3 Analysis ... 26 4.3.1 Data reduction ... 26 4.3.2 Data analysis ... 27 5. Results ... 27 5.1 Measurements of variables ... 27
5.2 Results of the proposed relationships ... 33
6. Discussion and Conclusion ... 35
6.1 Discussion... 35
6.2 Conclusions ... 38
6.3 Managerial implications ... 39
1. Introduction
The focus of health care information technology (IT) has been changing, from an emphasis on hardware, databases and systems architectures, to innovative uses of technology for facilitating communication and decision making, combined with a growing recognition of the importance of organizational and human factors (Pagliari et al., 2005). This focus has shifted because of the emerging need for efficiency in healthcare organizations. The trend of shortages of personnel in the healthcare sector is continuing, because of a lower supply of employees and an increasing demand on healthcare, which can cause future problems. Therefore, organizations in the healthcare industry are trying to find ways to avoid or limit these problems. In mostly commercial organizations, teleworking is already used as a solution to boost flexibility and ultimately efficiency. Teleworking is about the, how, why, where and when we work and this concept has never been as open to individual interpretation as today. Nowadays, employees are offered a wide range of work-life arrangements by their employer (Bond et al., 2005; Riedman et al., 2006). A broad range of organizations, are already offering laptops, smart-phones, tablets, flexible workplaces, and flexible work schedules, to be as flexible and efficient as possible. While teleworking is potentially favourable to both employers and employees, it has major implications for the organization and its employees (Bond et al., 2005). Benefits of teleworking can be divided into economic (productivity increase, reduction of costs) and non-economic benefits such as work-life improvements, higher job satisfaction and higher organizational commitment (Powell and Maineiro, 1999; Poelmans et al., 2003).
When looking at drawbacks, teleworking may lead to new sources of tension within an organization (Taskin and Edwards, 2007). Employees may be hesitant to adopt this new way of working because of a resistance to change, which may be caused by broken agreements, fear, communication breakdowns, lack of participation in the change and/or trust issues (Ford et al., 2008). Furthermore, telework may augment managerial uncertainty about the work being done correctly, because employees’ possibilities of acting in an untrustworthy way will increase (Peters and den Dulk, 2003). Organizational (cultural) controls may reduce these potential untrustworthy behaviours. These untrustworthy behaviours are mostly found in organizations which are characterized by low professionalism, less formalised rules and job evaluation systems and high centralisation in decision making. These types of organizations are therefore less suited for teleworking (Ruppel and Harrington, 1995). Managers should make an assessment of the risks of teleworking for their organization, and the costs involved in reducing those uncertainties. Moreover, teleworking may lead to social isolation among employees (Kurland and Bailey, 1999). Although many variations exist in the
teleworking practices across cultures, all implementation objectives share some resemblances, such as: enhancing organizational flexibility, lowering overhead costs improving flexible staffing practices, improving employees’ productivity and reducing commuting time for employees (Langa and Conradie, 2003).
In this thesis, factors influencing the intention to adopt teleworking at the individual level are examined. The intention to adopt, refers to an expressed desire to use or purchase a new product in the near future (Luarn and Lin, 2005). It relates to the employees state of mind before actual use or purchase behaviour will occur and is based on the information and perceptions the consumers or employees have at that time (Venkatesh and Davis, 2000). The research is done in a healthcare organization “Lentis”, which is currently examining the possibilities of teleworking and is in need for a framework with factors that help dealing with this innovation and that identifies factors which are important in influencing the intention to adopt teleworking at the individual level. In this study both theoretical and practical interests will be addressed. From a theoretical point of view, this study will contribute to the existing literature in multiple facets. Teleworking is relevant to examine as it is broader than technical innovations as it is not solely technology driven, but involves other dynamics, such as social, organizational and cultural implications (Mayo et al, 2009). Current teleworking literature is still not very informative about adoption factors as it is hampered by definitional problems, methodological weakness and lack of control of variables (Langa and Conradie, 2003). Most of the present literature in this field (Kurland and Bailey, 1999; Sullivan, 2003; Powell and Maineiro, 1999; Poelmans et al, 2003) focuses on organizational benefits and on the effects that teleworking has on the performance of individual employees and the way employees are affected by the teleworking phenomenon (Peréz et al, 2004). This thesis focuses on the intention of individuals to telework.It focuses on the different teleworking factors and their influence on adoption intentions at the individual level. Factors that influence individual adoption of teleworking are not yet identified in teleworking literature. Therefore, the purpose of this thesis is to contribute to the development of knowledge about teleworking by proposing and testing a model of teleworking adoption on the individual level. From literature, adoption model(s) will be taken and extended to fit the examined context. In current adoption literature, there is a focus on organizational and individual acceptance of single technologies, whilst not focusing on innovations comprehending more than technology (Venkatesh, 2012). By creating an adoption model for teleworking and therefore integrating social behaviours, technological factors, cultural factors, and organizational work factors into an adoption model, there is contributed to adoption literature. Furthermore, the identified factors potentially influencing the intention to adopt teleworking, will be tested in a healthcare setting. Most studies of telework, were predicated on home-based work (Shin et al, 2009). This study focuses mostly on
working on the go, which is defined as the opportunity for employees to perform work activities while not being at the office or at home, for instance whilst travelling (Shin et al, 2009). The gap in literature will be addressed with the following research question:
Which factors affect the intention to adopt teleworking among employees in a healthcare setting?
At Lentis, the term ‘WOW’ has been given to teleworking, which refers to the usage of tablets/laptops to work on the go and the flexibility of workplaces and work schedules. With flexibility of workplaces, the decoupling of work activity from a material workplace as the “office” is meant (Sullivan, 2003). With teleworking, individual work activities can be performed in house and outdoor and work meetings can be held in flexible places as well. In addition, more freedom will be created to flexible schedule work activities. Currently at Lentis, the contact moments with patients are registered in files, which can only be opened from within the office network. Because of the frequency of appointments with patients outside the office, the need arose for ambulatory teams to consult and update the files from outside the office. Ambulatory teams give treatment at home, at work or anywhere else to patients who do not live in an institution. In the future situation, which refers to ‘WOW’, a technical infrastructure will be created which makes it possible to open files of patients from outside the office. ‘WOW’ should avoid problems with inefficient travel and work practices. In addition, ‘WOW’ creates possibilities for updated patient information. ’WOW’ is defined by Lentis with four terms which are: ‘anytime and anywhere’, ‘my size fits me’, ‘manage your own work’ and ‘unlimited access and connectivity’. Furthermore, momentum for ‘WOW’ is created because evidence suggests that younger workers hold a “non-traditional” view of working life in comparison with older employees and are seeking for greater balance between their work and private life (Guest, 2002; Spangenberg & Lampert, 2009). Currently, Lentis has a staffing structure with an average age of fifty-two. This means that in combination with the aging, there will be a future shortage of personnel at Lentis because of the large outflow of employees in the next ten to fifteen years. This gap needs to be filled in by younger, graduated employees, which need to be attracted by ‘WOW’, as it stimulates the balance of working and private life and has a higher ICT component.
As a starting point, the literature section describes teleworking factors relevant on the individual level for explaining the attitude toward teleworking and the different adoption models. Thereafter, a decision for an adoption model is made. In addition, the variables proposed in the teleworking and adoption literature will be checked with initial interviews. Comparing variables found in literature with the variables mentioned by interviewees will perform this check. Variables will be used in this research, when they find support by various interviewees and can be supported by theory. On the
basis of this section, a research model will be presented. Next, the methodology part explains the research methods used. Then the analysis part will provide results and links theory to practice by describing phenomena and explaining them from a theoretical point of view. This will be followed by the conclusion section, where the research question will be answered. Then, contributions to the adoption models and teleworking literature are explained in the discussion section. Finally, the thesis concludes with a recommendation part, which elaborates on suggestions for future research.
2. Literature review
2.1 Teleworking
Teleworking represents a change in different dimensions, which are: technology, work organization, social behaviours and culture (Mayo et al, 2009). The changes in these factors are likely to influence individuals’ intention to adopt teleworking. In this section, variables which could be important in the context of individual teleworking adoption will be discussed based on the four factors of teleworking.
2.1.1 Technology
Telework can be regarded as a consequence of the development in ICT. Yet, this does not necessarily imply that it is a technologically superior way to carry out work. Rather it is a new way of organising work, asking for different social behaviour (Taskin and Edwards, 2007). With teleworking new technologies, such as tablets, smartphones and laptops are introduced to perform work activities. Research shows that technologies need to be functional and ask for minimal effort of individuals to stimulate acceptance (Davis, 1989; Venkatesh and Davis, 2000). These technologies such as the tablet, allows individuals for improved information exchange, multi-tasking, better organizing efforts and finally the tablet is proven to be ideal for displaying and analysing visual data in a healthcare setting (Garfield, 2006). Thus, technology represents one of the major components of teleworking because it enables workers to be in constant contact with the organization and their colleagues (Powell and Maineiro, 1999). As minimal effort and functionality are valued by individuals when using technologies, current work on the acceptance of these technologies confirms that perceived usefulness and perceived ease of use could be of great importance in influencing the attitude of individuals toward new technologies such as teleworking (Chau and Hu, 2002).
2.1.2 Work organization
In addition to a technology shift, the work organization will change when implementing teleworking. Technology is the main force but not the only force which brings about transformation to jobs and the organization and ways of organising work for greater responsiveness, flexibility and effectiveness (Poelmans et al, 2003). Moreover, teleworking brings work to workers rather than the other way
around. For that reason, teleworking involves the decoupling of work activity from one material workplace such as “the office” as well as from practices, schedules, scripts and prescribed working hours. With the change of work organization, time and costs are reduced by the reduction of commuting. This is one of the main reasons why employees choose telework (Kurland and Bailey, 1999). In addition, the time that individuals otherwise use for commuting or sitting in the office can be spend for their private life and potentially reduce work-family conflict (Powell and Maineiro, 1999; Poelmans et al., 2003). Furthermore, flexibility with planning work schedules will increase. People will be able to work anytime and anywhere, something that is reported to enhance productivity (Mann & Holdsworth, 2000; McCloskey and Igbaria, 2003). Teleworking increases self-reliance in scheduling particular tasks and augments control over the means of completing them (Gajendran and Harrison, 2007). These flexibility benefits provided by teleworking could temper negative psychological consequences of role stress experienced on the job and might increase acceptance of teleworking (Gajendran and Harrison, 2007). Teleworkers may not only be given a greater say in their spatial work conditions, they also get more decision-making authority and responsibility for control and adaptation of products and/or services. In addition, a feeling of enhanced autonomy among employees may be created when implementing teleworking because of the augmented flexibility in timing and execution of tasks (Gajendran and Harrison, 2007). Feelings of control will increase, as flexibility is stated to equal control (Gajendran and Harrison, 2007). These changes relate to the concept of empowerment and may enhance organizational commitment and job satisfaction (Klidas, 2001). A potential disadvantage in the change in work organization is that with teleworking more space is given to untrustworthy behaviour of employees as there is less control and more freedom in performing work activities (Powell and Maineiro, 1999).
2.1.3 Social behaviours
A consistent theme throughout literature reveals that social isolation is key in the limited adoption of teleworking (Ellison, 1999). This is not surprising as (informal) information dissemination, friendship formation and social networking typically occur in the workplace. With teleworking, colleagues are less available for social and task support at location and have a smaller role in influencing individuals’ attitudes (Ellison, 1999). Furthermore, the interaction of individuals with managers will flatten. As the transfer of managerial attitudes toward employees may stimulate individual acceptance, reduced interaction between managers and employees may inhibit teleworking acceptance (Ellison, 1999). In addition, reduced interaction between managers and employees may lead to role ambiguity, which refers to the lack of information about expectations concerning role performance (Olson, 1989). Teleworking may lead to role ambiguity which makes the concept more difficult to use (Olson, 1989). Furthermore, Hill et al (2003) indicate that workers involved in teleworking have more positive
perceptions about work/life balance and a have a greater feeling of personal success which may increase their perception of usefulness which in turn may increase adoption of teleworking among individuals. These are all effects of teleworking on the social behaviours of individuals and on the adoption of teleworking. As this thesis deals with the factors relevant for influencing individual intention to adoption and not with the direct effects of teleworking, other factor(s) should be identified. Factors which may influence the social behaviours of individuals and may stimulate the intention to adoption of teleworking are the attitudes and beliefs of colleagues/superiors about teleworking. This relates to the subjective norm concept, which expresses the perceived social or organizational pressure of an individual while intending the target behaviour which is needed (Taylor and Todd, 1995).
2.1.4 Culture
A change in culture will arise as a result of implementing teleworking, as the standard office culture cannot be integrated into a teleworking environment (Harrington and Santiago, 2006). Managers no longer are able to govern by “walking around” and observing employees. Control over employees has to be held in non-traditional ways, were it comes more to personnel controls (intrinsic motivation). Therefore, a culture of trusting has to be stimulated as it is seen as a vital prerequisite for telework (Ellison, 1999). Literature depicts that for teleworking to be accepted and flourish, a culture of trust has to be present, from both peers and managers of the teleworkers (Baruch, 2001; Mayo et al, 2009). Pre-existing work relationships between managers and teleworkers can augment trust and stimulate individuals to accept teleworking (Harrington and Santiago, 2006). Trust can be stimulated by providing procedures and rules to clarify responsibilities and rewards (Harrington and Santiago, 2006). In addition, trust can be stimulated among individuals by creating a culture in which employees are measured by results instead of attendance (Baruch, 2000).
2.2 Behavioral adoption models
When discussing the adoption models, the goal is to identify various theories and the factors which are commonly used in adoption studies and can be incorporated into the research model of this thesis. Both technological and behavioral adoption models are selected, as telework is not solely a technological innovation but also entails behavioral changes. The Technology Acceptance Model, Theory of Planned Behavior and the Unified Theory of Acceptance and Use of Technology are the models discussed in this section. In section 2.3 a decision for a research model will be made.
2.2.1 Technology acceptance model (TAM)
information technology. It replaces the attitudinal determinants with two beliefs: perceived usefulness and perceived ease of use. These are factors in the TAM model, influencing the attitude toward use (Chau and Hu, 2002). Perceived usefulness is defined as “the degree to which a person believes that using a certain technology will enhance his/her performance”.
The perceived ease of use can be defined as “the degree to which a person believes that using a certain technology will be free from effort”. Employees will assess the time and effort needed before the acceptance of technologies and innovations (Venkatesh, Thong & Xu, 2012).
TAM is found to be much easier to use by researchers than TRA and more powerful when considering the acceptance of information technology (Mathieson, 2001). While TAM is specific to information system usage, TRA is a general theory of human behaviour. The underlying supposition of TAM which depicts that perceived usefulness and perceived ease of use are always determinant for adoption decisions, is a limitation of the TAM model. TAM assumes usage as volitional, that is, there is no presence of barriers that may prevent individuals from using a technology if he or she chooses to do so. However, technology usage may be prevented by variables such as lack of time, expertise and money (Mathieson, 2001).
Figure 1: Technology Acceptance Model
2.2.2 Theory of planned behavior (TPB)
The theory of planned behaviour is one of the most widely used models which deals with explaining and predicting an individual’s behavioural intention and the acceptance of innovations. TPB is an attitude intention behaviour model, which states that people’s behaviour is directly influenced by their intention and moreover jointly determined by the attitude toward behaviour, the perceived behavioural control and the subjective norm. Behavioural intention is a measure of the strength of
an individual’s readiness about performing a certain behaviour (Wu and Chen, 2005). The more a behaviour depends on circumstances that are out of control or external for an individual, the less a behaviour is intentionally controllable. So, the TPB models next to the relationship between perceived behavioural control and intention also a direct relationship between perceived behavioural control and behaviour.
The attitude toward the behaviour is influenced by attitudinal beliefs of an individual, where normative beliefs influence the subjective norm and the power of control beliefs influences an individual’s perceived behavioural control (Meng, Chiuc and Chang, 2006). The attitude toward behaviour, refers to an individual’s favourable or unfavourable valuation about performing the target behaviour. Subjective norm expresses the perceived social or organizational pressure of an individual while intending the target behaviour which is asked for (Taylor and Todd, 1995). In other words, the subjective norm is the perception of an individual that most people who are important to him/her find that he/she should adopt the target behaviour. Perceived behavioural control refers to the difficulty or the convenience of performing the target behaviour. It concerns the beliefs about presence of control factors that can facilitate or impede the expected or required behaviour. Besides the given relationships, TPB is open to further expansion.
A limitation within the TPB model, is that it does not elaborate on the relationship between the belief structures and the antecedents for intention (attitude, subjective norm and perceived behaviour control). Within this model different belief structures are combined into one belief concept and therefore underlying dimensions of these belief structures may not be taken into account in determining the intention to adopt. These underlying dimensions may differ for different contexts and this is why TPB is hard to generalize among different settings. By decomposing the belief structures of the TPB model, a clearer relationship can be created and the model will become more appropriate and useful for practical situations and may be used in various settings (Wu and Chen, 2005).
Figure 2: Theory of Planned Behaviour model
In the Theory of Reasoned Action it has been proven that attitude influences behavioural intention (Ajzen en Fishbein, 1980). This relationship has received extensive empirical support (Chau and Hu, 2002; Wu and Chen, 2005; Jiang et al, 2000). In the research of Wu and Chen (2005), the attitudinal belief structure is decomposed into a variety of dimensions. The dimensions “ease of use”, and “perceived usefulness” are derived from the TAM model and “compatibility” is the other dimension for the attitudinal belief structure. The normative belief structure is decomposed into peer and superior influences. Control belief structure is decomposed into three dimensions: self-efficacy, technology facilitating conditions and resource facilitating conditions. Technology and resource facilitating conditions are integrated into the concept of controllability in various contexts (Pavlou, 2006). Controllability is defined as individual judgements about the availability of resources and opportunities to perform the behaviour (Pavlou, 2006)
2.2.3 Unified Theory of Acceptance and Use of Technology (UTAUT)
As can be derived from the last two paragraphs, a large stream of (technology) acceptation models exist. Researchers are confronted with many models and find that they must “pick and choose” constructs of the various models, or choose a “favoured model” and ignore some contributions of alternative adoption models. Therefore, Venkatesh et al (2003) reviewed these models and developed the UTAUT model, which is a comprehensive synthesis of these prior models, and should help to progress toward a unified view of user acceptance. UTAUT consists of four main concepts: effort expectancy, performance expectancy, social influence and the facilitating conditions. Effort
expectancy is the easiness associated with individuals’ use of a technology. Performance expectancy
refers to the degree in which benefits are created by using a technology in certain (work) activities.
because of the opinion of important others and the facilitating conditions refer to the perceptions of individuals about having the support and resource necessary to perform the behaviour. Age, gender, experience and voluntariness serve as moderators in this model (Venkatesh et al, 2003). Following UTAUT, effort expectancy, performance expectancy and social influence determine the intention to use. The facilitating conditions and the behavioural intention determine the actual use of a technology. Venkatesh et al. (2012) adapted the original UTAUT model and added a few dimensions. The first dimensions that is incorporated into UTAUT II is hedonic motivation and can be defined as the fun and pleasure that is derived from making use of a certain technology (Venkatesh, 2012). The term used for this concept by Thong et al. (2006) is perceived enjoyment and in this research it is found that it influences technology acceptance. Finally, in UTAUT II habit is introduced into technology acceptance theory. Habit is the degree that individuals perform behaviours automatically because of learning and the concept is operationalized in two different ways. Firstly, habit is seen as previous behaviour and secondly habit is measured as the degree to which a particular behaviour is seen as an automatic response for an individual. It seems that experience and habit are comparable, but there are at least two distinctions. The first distinction is that experience is needed to create habit, but on its own will not be sufficient. The second distinction is that by using a technology experience will grow which can result in differing levels of habit, dependent on the degree of familiarity and interaction with a particular technology (Venkatesh, 2012). Ajzen and Fishbein (2005) stated that feedback from earlier experiences have an influence on various beliefs and therefore future behaviour. In this light, habit can be seen as a perceptual concept that mirrors the results of earlier experiences.
Figure 3: Unified Theory of Acceptance and Use of Technology model II
2.3 Decision for an adoption model
After evaluation and comparison of the different adoption models it may be correct to say that there is quite some difference between the models in terms of their predictions about intention (Venkatesh, 2012). Based on the context and situational characteristics an adoption model will be chosen for this research. Davis (1989) used TRA for developing the TAM model, which is used for the individual acceptance of information technology. In the TAM model and the TRA model, attitude predicts intention and intention predicts behaviour. TAM does not include the subjective norm in comparison with TRA, because of the uncertain empirical and theoretical psychometric status (Davis, 1989). In comparison with the TRA model, the TAM model is more focused on analysing human behaviour for using information systems. As already mentioned, TAM is found to be much easier to use than TRA and more powerful when measuring the acceptance of an information technology (Mathieson, 2001).
The differences between the TAM and TPB model can be structured into three groups. The degree of generality, the social influence and the behavioural control (Davis et al. 1989). TAM always assumes that beliefs about perceived usefulness and ease of use are the primary determinants for the intention to use. In contrary, TPB assumes that beliefs are specific to a certain situation. To identify these beliefs, TPB needs a pilot study to identify reference group, control variables and relevant
outcomes for every situation, where TAM does not. This makes TPB more difficult to apply across diverse user contexts. Furthermore, social influence is not explicitly mentioned in the TAM model. Another difference lies in the behavioural control. The only behavioural control variable included in the TAM model is the perceived ease of use. Examining the items of this PEOU construct (Davis et al., 1989), it is clear that PEOU refers to the match of the users’ capabilities and the needed skills to use a certain technology. This corresponds to the internal factor of skill. In comparison with TPB and UTAUT the TAM model does not include external control factors. However, Li and Kishore (2006) concluded that researchers need to be cautious with interpreting results from studies using UTAUT scales as the different constructs have invariant true scores. Furthermore, in this thesis an adoption model has to be developed with those variables that specifically fit teleworking. This is because teleworking is not a change solely in technical factors, but also in culture, work organization and social behaviours (Mayo et al, 2009). Specific variables will be elaborated in this research and therefore TAM, which is simple in essence and has robust variables (perceived usefulness and perceived ease of use) is more suited than the extensive UTAUT model. In addition, factors from UTAUT such as price value are not relevant in the context of teleworking. Because of these reasons the UTAUT model is not chosen for this research. In spite of the limitations of the TAM model, the model is used in this research to measure the intention to adopt. The most important reasons are the following: the TAM model is the most commonly applied model of IT usage and adoption (Taylor and Todd, 1995). TAM has received considerable empirical support (Mathieson, 1991) and the two belief factors of the TAM model (perceived ease of use and perceived usefulness) are easy to understand. Furthermore, TAM is better suited to measure general levels of intention across a diverse population (Taylor and Todd, 1995).
As mentioned, the TAM model focuses solely on technology acceptance. As teleworking is not solely a technical innovation but also changes organizational practices and the entire way of working, the TAM model does not capture all factors influencing the intention to adopt teleworking. Therefore, next to technology factors the research model needs to be extended with cultural, social behavioural and work organizational factors relevant in the context of teleworking. These factors were discussed in section 2.1.
2.4
Initial interviews
Before the empirical continuance of this research the face validity of the model was tested with a series of semi-structured interviews. Semi-structured interviews are flexible and contain a list of topics to be explored. The model that was developed is said to have face validity when it “looks like” it is going to measure what it is supposed to measure (Anastasi, 1988). There might be important factors not included in the developed research model after the literature review of teleworking and adoption models in section 2.1 and 2.2. Therefore, a check of the model is held by semi-structured interviews, which may confirm current factors and/or explore missing factors in the research model. The interviewees were taken from various groups of the population, which are the different ambulatory teams, as the subjective norm may influence the attitudes towards teleworking within these teams and may limit different insights. Furthermore, a person from the board of Lentis was interviewed as he may held another view than the employees targeted to work by the concept of teleworking. In addition, the project leader of WOW was interviewed as he is supposed to hold a more holistic view.
The functions of the eight people interviewed are the following: two project leaders, a psychiatrist of the crisis service, a psychologist at Linis, a nurse at Welnis, a psychiatrist at Dignis, a member of the board of directors and a team leader.
A list with topics which were discussed during the interviews and an explanation for choosing these topics is presented in appendix 11.
Technology
As discussed in section 2.1 a change in technology will occur when teleworking is implemented. The perceived ease of use and perceived usefulness of these technologies are factors determining the intention to adopt teleworking. These variables are also considered important in a teleworking context:
‘When teleworking is not perceived as useful by employees, it has no point trying to implement it. Support from management alone is not enough; employees should have the capabilities and appreciate teleworking usage’ (Project leader)
Work organization
The work organizational factors, empowerment and perceived flexibility benefits, which are
‘In an organization such as Lentis with so many rules and bureaucracy, more freedom and
opportunities to perform work activities, will be received with pleasure by a lot of employees’ (project leader)
‘Because of the increase in autonomy which teleworking is supposed to offer, I would use the concept’ (psychiatrist crisis service)
‘I can’t wait to have access to patient files at different locations, this way I can finally report on site instead of hours later at the office’ (psychologist, Linis)
‘The chance to work everywhere and anywhere is very important for me, I believe flexibility gained with teleworking helps to optimize care for the patient and improves my own performance’(nurse, Welnis)
Social behaviours
As stated in section 2.1.3 social isolation inhibits telework adoption. Managers have less contact with their employees, which may limit their influence on employees’ attitudes. Section 2.2.3 indicates that social influence plays an important role in forming the behavioural intention of an individual.
Moreover, with teleworking social influence might decrease, as the contact between managers and employees is reduced. Still the subjective norm is believed to influence the behavioural intention to use teleworking, what is confirmed in an interview:
‘When people around me are positive about teleworking I will be more eager to use it, which is the
case with every kind of technology for me’ (psychiatrist for elderly, Dignis)
Culture
As described in the literature of teleworking in section 2.1.4, a culture of trust has to be present to let teleworking flourish. Trust in colleagues and top management is seen as a salient behavioural belief that affects employees’ attitude toward teleworking. Trust is confirmed to be important after some statements in the interviews:
‘Teleworking is all about trust, when managers and employees do not trust each other teleworking will not be adopted and is doomed to fail’ (member of the board of directors)
‘When my manager trusts me and I trust the managers, I see no reason why teleworking should fail’ (team leader)
Personal characteristics
The personal characteristic, personal innovativeness is a factor which is not derived from teleworking literature but is added to the research model as it determined as a missing factor after holding the initial interviews. Personal innovativeness is described in multiple studies as a determinant for technology acceptance (Agarwal and Prasad, 1998; Eastlick and Lotz, 1999). Personal innovativeness is said to determine perceived usefulness, perceived ease of use and the intention to adopt a certain technology (Lu et al, 2005). People with high personal innovativeness would be more likely to take advantage of a new technology and therefore may hold more positive feelings about teleworking (Lu et al, 2005). Innovative individuals are more likely to cope with high levels of uncertainty and therefore develop more positive intentions towards usage (Agarwal and Prasad, 1998). In various interviews with a member of the board, a psychiatrist and with the project leader management development, there was stated that expectations were present about a difference among different individuals and their attitude toward teleworking at Lentis. They think that different individuals have different traits which influence their attitude toward teleworking. There are innovative individuals, which they feel will accept teleworking earlier than others. This is supported by various literature (Eastlick and Lotz, 1999; Agarwal and Prasad, 1998; Lu et al, 2005). They stated the following:
‘Within Lentis different sorts of people exist, in general the younger people are more open to innovations and changes than their older colleagues’ (member of board of directors)
‘In my team there are people who are always first with adopting new technologies are given input to improvements’ (psychiatrist for elderly, Dignis).
‘There are some groups of individuals who are always first when it comes to bringing new ideas into the organization, I expect those employees to have a positive attitude towards teleworking as well’
(member of board of directors)
‘We are trying to attract young people to our organization as they fit the future of teleworking. The experience within Lentis is that younger people are more open to change’ (project leader)
3. Research model
After the section of the initial interviews where personal innovativeness was found as an important adoption factor, the research model can be presented. The different variables of the research model and the hypothesis will be described first. After which the conceptual model will be presented
3.1 TAM components
The dependent variable of the model will be the intention to use teleworking and is derived from the TAM model which is chosen as the basis of this research as described in section 2.3. The intention to adopt teleworking, refers to an expressed desire to use teleworking in the near future (Chau and Hu, 2002). The attitude toward using teleworking is another variable from the TAM model, it refers to an individual’s favourable or unfavourable assessment about performing the target behaviour and it is predicted to have a direct influence on the intention to use teleworking (Chau and Hu, 2002).
H1: The attitude toward usage has a positive influence on the intention to adopt teleworking 3.1.1 Technology
The two antecedents of attitude derived from the TAM model are perceived ease of use and
perceived usefulness.
Perceived ease of use refers to the degree to which a person believes that using teleworking will be
free from effort (Langa and Conradie, 2003). Lentis is offering very simple laptops to use in the beginning of WOW. After a certain amount of time more complex technologies as tablets, smartphones etcetera, will become available for employees of the ambulatory teams. The usage of these more advanced technologies is voluntary. It is expected that perceived ease of use of the technologies involved in telework will have a positive influence on the attitude toward telework.
H2: Perceived ease of use has a positive influence on the attitude toward using teleworking
Perceived usefulness refers the degree to which a person believes that using teleworking will
enhance his/her performance within an organizational context (Langa and Conradie, 2003). Expected is that teleworking increases employees scheduling flexibility, saves in cost and time spent in travelling, reduces travel related stress, and the ability of balancing private and work life will be augmented (Powel and Maineiro, 1999;Poelmans et al, 2003; Gajendran and Harrison, 2007). Furthermore, teleworking is expected to improve the reporting of the client as well as decrease administrative burdens. In TAM literature for various IT applications (Chau and Hu, 2002; Venkatesh, Thong and Xu, 2012), perceived usefulness is shown to influence the attitude toward using.
H3: Perceived usefulness has a positive influence on the attitude toward using teleworking
As mentioned, the TAM model focuses solely on technology acceptance. As teleworking is not solely a technical innovation but also changes work organization, social behaviours and culture, the TAM model does not capture all factors influencing the intention to adopt teleworking. Therefore, the model needs to be extended with other factors relevant for this context, which are elaborated in section 2.1 and section 2.4. The suggested extensions are described in the following section.
3.2 Extended model based on teleworking features and initial
interviews
3.2.1 Work organization
As mentioned in section 2.1.2, teleworking offers various flexibility benefits (Gajendran and Harrison, 2007; Powell and Maineiro, 1999; Poelmans et al, 2003). Perceived flexibility benefits refers to the degree to which a person can work anywhere and anytime. Following Nicolas et al (2008) a higher perceived flexibility benefit should increase the rate of adoption, as it positively influences attitudes of employees toward a certain technology. This is a rather important factor for the usage intentions of teleworking, as flexibility is one of the main advantages of teleworking in contrast with traditional ways of working (Poelmans et al, 2003).
H4: Perceived flexibility benefits have a positive influence on the attitude toward using teleworking Empowerment is creating and providing employees access to empowerment structures to perform
their work: opportunity, information, support, resources and formal and informal power. Both types of power should be associated with mastery and autonomy, instead of domination and control. Following the change literature, the increase in empowerment has a positive influence on job satisfaction and attitude toward work (Boonyarit et al 2010). Furthermore, empowerment is proved to positively influence organizational commitment and attitude towards a change (Coetsee, 1999). As telework is a change in the way of working (Taskin and Edwards, 2007), the commitment and the attitude toward this change may be directly influenced by empowerment. Furthermore, empowerment is believed to increase because of teleworking, as people get more autonomy to perform their work and receive the opportunity to perform their work at different locations at different times. Information will become accessible on more locations and more resources are given to employees to perform their work activities (Langa and Conradie, 2003). This increase in empowerment may be seen as an expected benefit for employees.
3.2.2 Social behaviours
The subjective norm is proven to be important in adoption decisions and is integrated into this research as it comprehends social influence factors and may influence teleworking adoption (Ellison, 2007; Wu and Chen, 2005). It expresses the perceived social or organizational pressure of an individual while intending the target behaviour that is needed. In other words, the subjective norm is the person’s perceptions that most people who are important to him/here believe he/she should adopt the target behaviour, which is using teleworking (Fishbein and Azjen, 1980). People tend to adjust their beliefs to confirm to a group. When a large portion of a group has a particular attitude, it is likely that an individual will adopt that attitude as well (Asch, 1951).
H6: The subjective norm has a positive influence on the attitude toward using teleworking
3.2.3 Culture
Section 2.1.4 describes that a culture of trust is an important factor in teleworking adoption (Mayo et al, 2009). Trust is seen as an important aspect of culture in teleworking adoption. In a model of Gefen et al. (2003) there was explicitly indicated that trust has a direct influence with the intention to adopt. Furthermore, trust is considered an antecedent of attitude. When top management and other colleagues throughout the organization are trustworthy, it is more likely that employees will gain benefits and avoid possible risks when adopting teleworking. Trust in colleagues and top management is viewed as a salient behavioural belief that directly affects employees’ attitude toward using. As the cost- benefit paradigm influences people’s attitudinal beliefs and outcome judgements, trust might be a direct factor influencing the attitude toward using (Wu and Chen, 2005). When employees initially trust their peers/team leaders and have the idea that teleworking beneficially influences their job performance, they will believe that teleworking is useful (Wu and Chen, 2005).
H7: Trust has a positive influence on the attitude toward using teleworking 3.2.4 Personal characteristics
Personal innovativeness may influence the intention to use teleworking and can be defined as the
willingness of an individual to try out any new technology. Teleworking is expected to be accepted earlier by progressive individuals with high personal innovativeness, in comparison with persons with less personal innovativeness. Eastlick and Lotz (1999) developed a theoretical model that linked personal innovativeness traits with adoption intentions through an attitudinal concept. They found that persons, who have high personal innovativeness, were more likely to have a higher intention to adopt innovations. Various literature supported these findings and found that personal
innovativeness has a direct influence on attitude toward using a technology (Agarwal and Prasad, 1998; Lu et al, 2005).
H8: Personal innovativeness has a positive influence on the attitude toward using teleworking
3.3 Control variables
Gender, voluntariness, years at organization, years in function, team/care group, age and experience with teleworking are added as control variables in this research. It is believed that user intention will be influenced by interfering factors as gender, age, experience and voluntariness (Venkatesh et al, 2003; Moore and Benbasat, 1991). Research on gender indicates that men tend to be highly task-oriented and more willing to accept technologies, which may improve results and benefits (Venkatesh et al. 2003). In addition, females potentially have higher anxiety for new technologies and consequently men may adopt teleworking earlier (Igbaria and Chakrabarti, 1990). Age is proven to play a role in technology acceptance, as younger people are more open to new ideas and earlier in technology acceptance (Venkatesh et al, 2003). Furthermore, experience is said to positively influence innovation adoption as it positively influences perceived ease of use and perceived usefulness (Venkatesh et al, 2003). Voluntariness is likely to influence the intention to adopt teleworking, when employees are mandated to work following the concept of teleworking, the intention to adopt may increase. Years in function, years at organization and team/care group may influence the adoption decision, as some teams/care groups are better suitable for teleworking than other teams/care groups. In addition, when working for a long time in a certain organization and function, certain rustiness can evolve because of rooted routines and norms. This may reduce the readiness to change and therefore limit teleworking adoption (Lewin, 1954).
Perceived usefulness Perceived ease of use Perceived flexibility benefits Empowerment Subjective norm Trust Personal innovativeness Attitude toward using teleworking Intention to adopt teleworking Control variables *Gender *Age *Experience *Voluntariness *Years in function *Years at organization *Team/care group + + + + + + + + Technology Work organization Social behaviours Culture Personal characteristics
Figure 4: The research model
4. Research methodology
4.1 Measures
Items for an online questionnaire were developed in order to test the hypotheses that were proposed in section 3. For each construct, existing items sets from the extant literature were used, meaning that this study was performed making use of widely used multi-item scales that have been proven to be reliable and valid. However, the existing items had to be translated into Dutch and slightly modified to fit a teleworking setting.
The items of the constructs perceived ease of use and perceived usefulness of the TAM model were adopted from Venkatesh (2000) and slightly modified to fit the teleworking context. Intention to adopt was also adopted from Venkatesh (2000) but needed more adaption as some items needed
adopted from Bhattacherjee (2000) as he used more items for this construct than Venkatesh (2000) who used a rather limited scope of items. In addition, the items that Venkatesh (2000) used for attitude toward using are rather similar to the items used for the intention to adopt. The perceived flexibility benefits items were adopted from Nicolás et al (2008) as they also used these items in a technology acceptance study. The empowerment items were adopted from Boonyarit et al. (2010) and Gill et al. (2012). The subjective norm items were used from Bhattacherjee (2000) and Taylor and Todd (1995) which used these items in adoption intentions research. Trust was adapted from Gefen et al. (2003) and Goodwin et al. (2011). Goodwin et al. (2011) used scales for interpersonal trust which are examined in this teleworking setting and Gefen et al. (2003) integrated trust in a TAM setting. For the last construct personal innovativeness items were used from Agarwal and Prasad (1995). The original items are presented in appendix 1 and the translations in appendix 2. Appendix 3 gives an overview of these original constructs and translations in one table.
All items were measured with a 5-points Likert scale that ranged from 1 corresponding with ’totally disagree’ to 5 corresponding to ‘totally agree’. The initial constructs and items were tested using a pilot. This pilot was done with 6 people working in one of the ambulatory teams. This ambulatory team was selected as various functions, age groups and levels of experience were present among the individuals in this team. In addition, male and females are represented more or less equally in this team. Furthermore, when looking at practical implications, this was also one of the few teams available in the holiday period. After this pilot the following adjustments were made in the items of empowerment. Item 7 (appendix 1) was added in the questionnaire after deleting another item of empowerment that was not understood by the participants of the pilot. The others items of empowerment needed a little adjustments in wording. The rest of the constructs and items were clear to the employees involved in the pilot.
4.2 Sample
The research was conducted at the mental healthcare organization Lentis. The population consisted of employees working in ambulatory teams, as they were the first teams starting to work following the concept of teleworking. These teams did not yet started working following the concept of teleworking during this research but would start in November of 2012. It was necessary to spread the questionnaires before the start of teleworking, as this research measures the intention to adopt teleworking and not the actual adoption. In the preamble of spreading the questionnaire, information about teleworking was spread throughout the organization by using Yammer, which is an internal social media program for Lentis employees comparable to Twitter. In addition, the weekly intern paper of Lentis was used to distribute information about the research and on behalf of the
board of directors mails were send throughout the organization to confirm the importance of this research and to communicate what ‘WOW’ will exactly encompass for them. In the questionnaire an introduction on teleworking was presented once more (appendix 1).
The e-mail with the link to the questionnaire was sent in July 2012 to a sample of 196 employees out of the population of 343 employees. There was a large group of employees who already participated in another research about teleworking and where therefore excluded from the sample. The sample was selected from the population with convenience sampling, to get a gross estimate of the results without incurring the time and cost to select a random sample (Lunsford, 1995). Initially the mail was sent to 11 team leaders, who in turn distributed the e-mail with link to the 196 employees in their teams. They indicated the importance of filling in the questionnaire to their employees and after calling all the individual team leaders they even brought the research and questionnaire up in their team meetings. The respondents were given a period of five weeks to complete the questionnaire. During this period the team leaders sent one reminder, which was already drawn for them, to all the employees. A long period was given for filling in the questionnaire due to the holiday period of employees. In total 93 questionnaires were filled in. Only respondents who completed the questionnaire were included in the sample, so no missing values were present. This is a response rate of 47, 5%. Based on the descriptives shown in table 1 and 2, there is no reason to suspect that the non-response is select and bias exists. The descriptives such as age, gender, years in organization and team/care group are almost evenly distributed in the response rate as compared to the population as can be seen in table 1. Therefore, there is no reason to assume that select groups of people of the population are causing a select non-response.
Table 1: Age, gender, years in organization and team/care group in population and response group
Population Respons group
Age (mean) 46 43 Gender Man: 45% Women: 55% Man: 44,1% Women: 55,9% Years in organization (approximately) 11 12 Team/care group (percent) Linis: 49,6% Welnis/crisis service: 7,6% Bing: 3,2% Dignis/elderly psychiatry: 31,7% Team 290: 7,7% Linis: 59,1% Welnis/crisis service: 11,8% Bing: 3,2% Dignis/elderly psychiatry: 20,4% Team 290: 5,4
Table 2: Descriptives of respondents (control variables)
Options Response
(n = 93)
Percent (%) Mean Std dev.
Gender Man Woman 41 52 44,1 55,9 1,56 0,499 Age Much experience Voluntariness Years in function Years at organization Team/care group 21-30 31-40 41-50 51-60 61-70 Totally disagree Disagree Neutral Agree Totally agree Totally disagree Disagree Neutral Agree Totally agree 0-5 6-10 11-15 16-20 More than 20 0-5 6-10 11-15 16-20 More than 20 Linis Welnis/crisis service Bing
Dignis/ elderly psychiatry Team 290 4 22 29 32 6 8 17 31 22 15 9 44 36 3 1 52 22 6 6 7 15 17 19 11 31 55 11 3 19 5 4,3 23,4 30,9 34,4 6,4 8,6 18,3 33,3 23,7 16,3 9,7 47,3 38,7 3,2 1,1 55,3 23,4 6,4 6,4 7,4 16 18,1 20,2 11,7 33,0 59,1 11,8 3,2 20,4 5,4 3,15 3,22 2,41 1,86 3,28 - 0,999 1,044 0,709 1,248 1,492 -
4.3 Analysis
4.3.1 Data reductionThe objective of this study is to test all the hypotheses in order to find the constructs that significantly influence attitude toward using teleworking and the intention to adopt teleworking. To test the hypotheses and find if the constructs are of influence some preconditions had to be met. First, the validity of the items was tested. The various items were put into a factor analysis to see if the items loaded on the correct constructs. Before performing a factor analysis first the Kaiser-Meyer-Olking (KMO) measure of sampling adequacy and a Bartlett’s Test of Sphericity was done to
see whether a factor analysis was useful. The factor analysis was then performed with Varimax rotation that maximizes the sum of the variances of the squared loadings. After the factor analysis a Cronbach’s alpha analysis was performed to measure the reliability and internal consistency of the constructs. This analyses identifies if items can be excluded and determines with which of the loading items the construct can be measured most reliably.
4.3.2 Data analysis
Finally, regression analyses were performed to test the hypotheses and to define which factors influence the intention towards using teleworking. This analyses identifies tolerance scores and the VIF to check for multicollinearity, which refers to a situation in which two or more explanatory variables in the multiple regression analysis are highly linearly related. The mathematical model used is as follows:
Attitude toward using teleworking = f (perceived usefulness, perceived ease of use, perceived flexibility benefits, empowerment, subjective norm, trust, personal innovativeness)
The items that represent each of the constructs are provided in appendix 1. Appendix 2 gives an overview of the original English wording of these items.
5. Results
5.1 Measurements of variables
Factor analyses were performed for the independent items and a separate analysis for the dependent items and the control items. A separate analysis was performed for the dependent variables as they did not load correctly in a factor analysis with the independent variables. Items of the dependent variables loaded on the independent variables. The KMO and Bartlett’s test showed that it was justifiable to perform the factor analysis for the research model for all items of the independent variables together (0.799) and for the items of the dependent variables (0.917). In the factor analysis the requirement was used that the different items should load on a component with 0.400 otherwise the decision was made to delete that item(s) (Gorsuch, 1983). If the items loaded on more factors there was seen which loadings were higher. Furthermore, the Cronbach’s Alpha was used to check for reliability and internal consistency. Most researchers look for Cronbach’s Alpha exceeding 0.7(Nunnally, 1978). Therefore, after the factor analysis items were deleted if that helped to raise internal consistency from a value lower than 0.7 to a value higher than 0.7.
In the factor analysis, only items 1, 3 and 4 of the construct perceived usefulness loaded, items 2 and 5 were excluded (table 3). The items 1, 3 and 4 of this construct generated α= 0.795 (table 5), which justified the results of the factor analysis. The items 7 and 8 of the construct empowerment loaded on the same component as perceived usefulness while items 6 and 9 loaded on the same component as perceived ease of use. Items 6 and 9 showed a very low internal consistency together (α=0.423). Items 7 and 8 of empowerment have a α=0.622. These items of empowerment load on the constructs
perceived usefulness and perceived ease of use as there is overlap between the constructs.
Empowerment deals with opportunities, autonomy and formal and informal power which partially
overlap with perceived usefulness as these characteristics of empowerment are seen as performance benefits for this construct. In addition, the support and autonomy component of empowerment overlap with perceived ease of use, as that construct deals with the ease of using an innovation which is facilitated by support. The construct of empowerment is excluded from the overlap with perceived
ease of use and perceived usefulness, where the residue value will represent the construct empowerment. This will be done by regression analysis and thereafter a factor analysis and
Cronbach’s alpha is performed for the residual values from the items to see if they load correctly and a variable can be formed. After performing this analysis all new items did load correctly (appendix 5) with a α= 0.665 (appendix 6 and table 5). This is below the 0.7 but the scale was maintained because this value is still acceptable following Nardo et al, (2008) and the factor analyses loaded higher than 0.4. In addition, another factor analysis was performed with the constructs perceived usefulness, perceived ease of use and empowerment (4 new items consisting out the residual values), where the three constructs loaded correctly (appendix 7). There is enough support to handle perceived
usefulness and perceived ease of use as different from empowerment after correcting empowerment
for the overlap with perceived ease of use and perceived usefulness. Literature also supports this, Boonyarit et al (2010) state that empowerment deals with opportunities, information, support, resources and formal and informal power for employees while perceived usefulness deals with the performance benefits that teleworking may create for employees (Langa and Conradie, 2003). For the subjective norm, item 21 and 23 loaded correctly in the factor analyses (table 3) with an α=0.557 (table 5), item 22 was excluded as it did not load correctly in the factor analyses. Item 10 and 11 of perceived flexibility benefits loaded correctly (table 3), item 12 was excluded. The α=0.715 for these two items of perceived flexibility benefits, which is higher than the alpha when using all three items as shown in table 5. For the construct perceived ease of use item 18, 19 and 20 loaded correctly (table 3), the α= 0.844 for these three items (table 5). The items of personal innovativeness and trust all loaded correctly (table 3) and their reliabilities were sufficient (table 5). The adjusted
factor analysis, with exclusion of not loading items, from the independent variables is presented in table 3. The original factor analysis is presented in appendix 4.
Table 3: Factor analysis of independent variables without excluded items
Component Items 1 2 3 4 5 6 Perceived usefulness 1 -.021 .757 .237 .156 .085 .129 Perceived usefulness 3 -.033 .763 .215 .303 .147 .103 Perceived usefulness 4 .013 .564 .171 .344 .373 -.200 Empowerment 7 .182 .661 .084 .054 .231 .007 Empowerment 8 .437 .713 .044 -.155 -.091 -.012
Perceived flexibility benefits 10 .255 .239 .088 .171 .725 .101
Perceived flexibility benefits 11 .154 .142 -.016 .037 .868 .166
Personal innovativeness 13 .783 .052 .199 .178 .165 -.099
Personal innovativeness 14 .803 .045 .110 .208 .202 .028
Personal innovativeness 15 .792 .127 -.086 .106 -.017 .080
Personal innovativeness 16 .748 .107 .235 .355 .178 .088
Perceived ease of use 18 .511 .115 .029 .551 .283 .073
Perceived ease of use 19 .436 .177 .055 .757 .119 .025
Perceived ease of use 20 .272 .174 -.020 .825 .025 .097
Subjective norm 21 .053 .231 -.035 -.077 .159 .811
Subjective norm 23 .015 -.129 .199 .214 .066 .796
Trust 24 .164 .204 .907 .041 -.020 .104
Trust 25 .060 .197 .892 .050 -.028 .069
Trust 26 .080 .117 .914 -.017 .144 -.004
The factor analysis for the two dependent variables, the intention to adopt teleworking and attitude toward using teleworking, resulted in the following results. All the items loaded on the same component. Therefore, the two variables which are suggested to be different in literature (Luarn and Lin, 2005; Wu and Chen, 2005) and therefore treated independently in the research model and literature review, are integrated into one variable after this factor analysis. This variable will be called the attitude towards using teleworking and is part of the intention to adopt teleworking. The α=0.923 confirms the internal consistency between the two variables and supports the decision for integrating attitude toward using teleworking and intention to adopt teleworking. Therefore, a
regression analysis and test will not be necessary for hypothesis 1, as the intention to adopt teleworking is equal to and integrated into the attitude toward using teleworking. The factor analysis of the dependent variables is presented in table 4 and the correlation, mean and standard deviation of this new formed variable is presented in table 6.
Table 4: Factor analysis of dependent variables
Component
Items 1
Attitude toward using teleworking 27 .864 Attitude toward using teleworking 29 .689 Attitude toward using teleworking 28 .901 Attitude toward using teleworking 30 .917 Intention to adopt teleworking 31 .890 Intention to adopt teleworking 32 .839 Intention to adopt teleworking 33 .772
For the control variables voluntariness and experience a factor analysis is performed because they consist out of more than one item. For determining the Cronbach’s alpha, item 34 and 36 are recoded as they were asked in a negative way as opposed to item 35. For voluntariness all items (34, 35, 36) loaded correctly with a α= 0.721 while for experience only item 44 was excluded as it not loaded correctly. These items (41, 42, 43) of experience gave a α= 0.809. The adjusted factor analysis, with exclusion of non-loading items, from these two control variables is presented in appendix 8 and the Cronbach’s alpha of the different constructs is presented in table 5.
Table 5: Internal consistency of the scales before and after adjustment (Cronbach’s alpha) and the explained variance after adjustment
Construct Items Alpha 1 Items Alpha 2 Explained
variance
Perceived Usefulness Perceived ease of use Perceived flexibility benefits
5 4 3 0.740 0.851 0.532 3 3 2 0.795 0.844 0.715 37,6% 46,5% 32,1% Empowerment Subjective norm Trust Personal innovativeness
Attitude toward using teleworking Intention to adopt teleworking Voluntariness Experience 4 3 3 4 4 3 3 4 0.631 0.501 0.926 0.857 0.884 0.847 0.721 0.703 4 2 3 4 7 7 3 3 0.665 0.557 0.926 0.857 0.923 0.923 0.721 0.809 40,0% 25,1% 11,1% 31,7% - - 11,3% 11,3%
